1964: Mixed blessing
July 6, 1999
by Leslie Goff
(IDG) -- A single product announcement in 1964 ushered in a new era of computing. The new product was both a help and a hindrance to the IT profession, a strategic advantage that over 35 years would have unforeseen consequences.
It would simplify computer purchasing decisions but complicate computer operations. It would increase the amount of work that could be accomplished but impede individual productivity. And it would create new demand for information technology professionals but diminish the supply of people who could do the job.
On Tuesday, April 7, 1964, IBM's then-CEO, Thomas J. Watson Jr., unveiled the S/360 product line -- the first family of compatible computers. The six computers and 44 peripherals were all designed to provide upward compatibility, to work together and to run the same software. Plus, for the first time, the company announced software with the hardware: the OS/360 operating system, the PL/1 programming language and a compiler.
Today, the S/360 is hailed as one of the top 100 technological developments of the 20th century. It pioneered the concept of compatibility, a strategic victory for IT shops (and for IBM). It provided greater speed and power, and it introduced multiprocessing, enabling several jobs to run at once. But the new architecture, based on eight-bit memory, may have been both a blessing and a curse.
"The S/360 was the single biggest mistake in computing history," says Barry Gordon, now retired, who was at IBM at the time. "April 7 is a day that will live in infamy because it's the day the computing world became saddled with the hexadecimal numbering system."
Gordon, who later would manage the S/360 Model 40 product line, contends that the switch to a hardware architecture that used a base-16 numbering system rather than a base-10 system made computing far more complex than necessary and had a reverberating effect on the industry and the IT profession.
"Base-16 has been a source of errors and difficulty and will be forever because we are basically base-10 people," he says. "We've been stuck with it for 35 years, and it's a disaster."
With IBM's 650 or 705 base-10 machines, anyone could learn to program. Programming in base-16, Gordon says, required mathematically inclined individuals who could think more abstractly. The increased complexity, coupled with the intricacies of the operating system, "closed the door to a lot of people getting into programming," he says.
It also changed the nature of the job. Instead of a programmer sitting at a console and interacting directly with the hardware, now he handed his card set off to an operator and returned later to pick up the results.
That distancing of the programmer from the computer was inhibiting, says Ed Seidman, now a senior systems analyst at Abbott Laboratories in Chicago. In 1964, he used an early S/360 to design radar for ballistics tracking systems for North American Aviation in California.
The S/360 also prevented programmers from changing their code on the fly. If results weren't right the first time, they had to resubmit the job, wait again and eventually take the results back to an impatient user. That created a gulf between the user and the hardware that would only widen until the advent of the PC.
But on that day in April 1964, the user community was swept away. The idea that they wouldn't need to rewrite all of their software every time they needed new hardware meant a huge savings of time and money. What Fortune magazine had called IBM's "$5 billion gamble" yielded a huge payoff: Within four weeks, IBM had 1,000 orders for the computers and associated peripherals. Two years later the company had amassed 9,000 orders; by 1967, IBM's sales and leasing revenue had surpassed $5 billion.
And the S/360's multiprocessing capabilities made companies more productive. Although individuals had to wait longer for their programs, overall projects could be completed more quickly, and more balls could be in the air at once. Seidman remembers that jobs proliferated and programmers' salaries skyrocketed.
Once the bugs were worked out, the S/360 was a performer, and it put IBM's competitors in a reactive mode.
Even the skeptical Gordon concedes, "Aside from base-16, the machines were not bad."
Goff is a frequent contributor to Computerworld. Contact her at firstname.lastname@example.org.
You can tell a lot about an OS from its games
RELATED IDG.net STORIES:
The S/360 learning process
|Back to the top||
© 2001 Cable News Network. All Rights Reserved.|
Terms under which this service is provided to you.
Read our privacy guidelines.